Perplexity AI announced what it calls the first hybrid local-server inference orchestrator at Computex 2026. The system is designed to automatically route AI tasks between a user’s local device and cloud-based frontier models without requiring the user to decide in advance. The feature is expected come to Perplexity Computer in July 2026.

What is Hybrid Agentic Inference?

To understand what Perplexity built, it helps to understand the three-way tension that AI systems face.

Accuracy demands the most capable models, which are expensive to run. Privacy demands that some data never leave the device. Cost and energy efficiency demand that you don’t spend a frontier model’s compute on tasks a smaller model can handle.

That routing layer is what Perplexity calls hybrid agentic inference.

A compact AI model runs locally on the user’s device. This local model evaluates each incoming task or subtask. It determines whether the task involves sensitive data, whether it requires heavy computation, or whether it can be handled entirely on-device. Based on that evaluation, work is either kept local or sent to a frontier model in the cloud.

Perplexity describes this local model as deciding “when sensitive data should also be kept locally.” The system is designed to ask for user permission before sending sensitive tasks to the cloud. That design addresses a specific concern enterprises have about agentic AI: data governance — knowing where data goes and who controls that decision.

Examples of data the system is intended to keep local include financial records, health information, and personal files. Work that requires a frontier model’s full capability runs on the server. Most real tasks are a mix, so the system splits them and coordinates the parts.

How It Fits into Perplexity Computer

Perplexity Computer is the company’s cloud-based multi-model agentic product, launched in February 2026. It originally ran entirely in the cloud on the Perplexity Max subscription tier ($200/month).